Exacerbation of substrate toxicity by IPTG in Escherichia coli BL21(DE3) carrying a synthetic metabolic pathway

Abstract

Background: Heterologous expression systems based on promoters inducible with isopropyl-β-D-1-thiogalactopyranoside (IPTG), e.g., Escherichia coli BL21(DE3) and cognate LacI(Q)/P(lacUV5)-T7 vectors, are commonly used for production of recombinant proteins and metabolic pathways. The applicability of such cell factories is limited by the complex physiological burden imposed by overexpression of the exogenous genes during a bioprocess. This burden originates from a combination of stresses that may include competition for the expression machinery, side-reactions due to the activity of the recombinant proteins, or the toxicity of their substrates, products and intermediates. However, the physiological impact of IPTG-induced conditional expression on the recombinant host under such harsh conditions is often overlooked.

Results: The physiological responses to IPTG of the E. coli BL21(DE3) strain and three different recombinants carrying a synthetic metabolic pathway for biodegradation of the toxic anthropogenic pollutant 1,2,3-trichloropropane (TCP) were investigated using plating, flow cytometry, and electron microscopy. Collected data revealed unexpected negative synergistic effect of inducer of the expression system and toxic substrate resulting in pronounced physiological stress. Replacing IPTG with the natural sugar effector lactose greatly reduced such stress, demonstrating that the effect was due to the original inducer's chemical properties.

Conclusions: IPTG is not an innocuous inducer; instead, it exacerbates the toxicity of haloalkane substrate and causes appreciable damage to the E. coli BL21(DE3) host, which is already bearing a metabolic burden due to its content of plasmids carrying the genes of the synthetic metabolic pathway. The concentration of IPTG can be effectively tuned to mitigate this negative effect. Importantly, we show that induction with lactose, the natural inducer of P lac , dramatically lightens the burden without reducing the efficiency of the synthetic TCP degradation pathway. This suggests that lactose may be a better inducer than IPTG for the expression of heterologous pathways in E. coli BL21(DE3).

Scheme 1

Five-step biotransformation of 1,2,3-trichloropropane into glycerol by the enzymes of the synthetic biodegradation pathway. The pathway consists of haloalkane dehalogenase DhaA from Rhodococcus rhodochrous NCIMB 13064 [33] with the haloalcohol dehalogenase HheC and epoxide hydrolase EchA from Agrobacterium radiobacter AD1 [34, 35]. Computer-assisted protein engineering was used to improve activity of the haloalkane dehalogenase towards 1,2,3-trichloropropane, leading to the development of the 32-fold more active and 26-fold more efficient mutant DhaA31 [31]. Formed glycerol can be utilized in central catabolic pathways of the host cell

Fig. 1

Biotransformation of TCP catalysed by different Escherichia coli BL21(DE3) recombinants. a Control strain carrying the empty pCDF and pETDuet plasmids with streptomycin and ampicillin resistance marker genes, respectively. The blue arrows indicate individual T7 promoters. b The degrader degWT, which carries the haloalkane dehalogenase gene (dhaA) on pCDF and the haloalcohol dehalogenase (hheC) and epoxide hydrolase (echA) genes on pETDuet. c The degrader deg31, which carries the haloalkane dehalogenase mutant (dhaA31) gene on pCDF and two remaining genes of the degradation pathway on pETDuet. d The degrader deg31opt, which carries the dhaA31 gene on pCDF and the two remaining genes of the degradation pathway on pACYC along with a chloramphenicol marker gene. The relative ratios of the TCP pathway genes produced by the degraders degWT, deg31, and deg31opt are 0.24:0.36:0.40, 0.14:0.41:0.45 and 0.63:0.16:0.21, respectively; the corresponding theoretical conversions of TCP into glycerol (GLY) are 35, 68, and 44 %, respectively. Error bars represent standard deviations calculated from three independent experiments. Theoretical concentrations of GLY were calculated from experimentally determined concentrations of TCP and intermediates. Sm ^R streptomycin marker gene; Amp ^R ampicillin marker gene; Cm ^R chloramphenicol marker gene; DCP 2,3-dichloropropan-1-ol; ECH epichlorohydrin; CPD 3-chloropropane-1,2-diol; GDL glycidol. Note that the green line representing ECH is not visible because this intermediate does not accumulate at detectable levels during the experiment

Fig. 2

Effects of metabolic burden and TCP toxicity on physiological parameters of Escherichia coli BL21(DE3) cells and three recombinants expressing the synthetic metabolic pathway. a Viability of cells non-induced or pre-induced with IPTG determined by plating before incubation in phosphate buffer. The effects of metabolic burden stemming from the presence of plasmids, pre-induction with 0.2 mM IPTG, and expression of the synthetic pathway are indicated by coloured arrows. Asterisks denote significance in decrease of cell count caused by each of three effects at either P < 0.05 (*) or P < 0.01 (**) when compared with preceding condition. b The percentage of surviving cells (upper graph) and the corresponding physiological parameters determined by flow cytometry (lower graph) after incubation in buffer with or without 2 mM TCP. The separate effects of TCP, IPTG, and the exacerbation of TCP toxicity in cells pre-induced with IPTG are indicated by coloured arrows. Asterisks denote significant difference in the decrease of cell count caused by each of three effects at P < 0.01 when compared with preceding condition. Physiological parameters including membrane permeability, formation of reactive oxygen species (ROS), and membrane depolarization were evaluated by staining the cells with appropriate dyes as explained in the Methods section. Error bars represent standard deviations calculated from at least five independent experiments. CFU colony forming units; host-p E. coli BL21(DE3) without plasmids; host E. coli BL21(DE3) with the empty pETDuet and pCDF plasmids

Fig. 3

Transmission electron microscopy of Escherichia coli deg31 cells and corresponding histograms showing the physiological state of populations stained with selected fluorescent dyes. a Non-induced cells incubated in phosphate buffer. b Non-induced cells incubated in phosphate buffer with 2 mM TCP. c Cells pre-induced with 0.2 mM IPTG incubated in phosphate buffer. d Cells pre-induced with 0.2 mM IPTG and incubated in phosphate buffer with 2 mM TCP. Black arrows indicate bodies that presumably consist of overexpressed heterologous proteins, grey arrows indicate separations of the inner and outer cell membranes, and white arrows indicate dead or dying cells

Fig. 4

Viability of Escherichia coli deg31 and host control strains after pre-induction with diverse concentrations of IPTG or 1 mM lactose, before and after incubation with TCP. a Viability of deg31 and E. coli BL21(DE3) with the empty pETDuet and pCDF plasmids as determined by plating of cells pre-induced with different concentrations of IPTG or 1 mM lactose (red columns) before incubation in phosphate buffer with TCP. b Viability of cells after incubation in buffer with 2 mM TCP. Asterisks denote significantly higher (at P < 0.05) cell count of deg31 pre-induced with 1 mM lactose when compared with the count of cells pre-induced with the lowest tested concentration of IPTG (0.01 mM). c Fraction of surviving cells calculated as the difference in the CFUs.ml⁻¹.OD₆₀₀⁻¹ before and after incubation with TCP. Error bars represent standard deviations calculated from at least four independent experiments. CFU colony forming units

Fig. 5

Summarized effects of IPTG concentration on gene expression levels, pathway output, and cell viability in pre-induced Escherichia coli deg31 cells. Viability was determined by plating pre-induced deg31 cells resuspended in phosphate buffer before incubation with TCP. Pathway output was expressed as the theoretical conversion of TCP into glycerol at the end of 5 h degradation experiments with pre-induced, resting deg31 cells (see also Additional file 1: Fig. S5). The content of TCP pathway enzymes was estimated by analyzing cell-free extracts obtained from pre-induced cells by sodium dodecyl sulfate polyacrylamide gel electrophoresis (Additional file 1: Fig. S7 and Table S1). Two gels were analysed by densitometry and mean values are shown. Error bars represent standard deviations calculated from three independent experiments. Values determined for deg31 pre-induced with 1 mM lactose are indicated by squares